ZX Appeal

JAN/90

vancouver sinclair
users gr oup

пехт meeting

ЕЕ ано, 2

BITS PIECES оне 2

MEETING MINUTES..............- 2

HARVEY Gist ынады нанына. 3

January 12, 1990 АНАМА а нан 5
1000/2068 РЕОС...................... 7

SOLDERING ТИР 8

Е mti CURVES-2068 РКОС.............. 9
SINCLAIR STORY PART 11...12

newsletter put out by the

Vancouver Sinclair Users Group.

For more information on the

group and ZXAppeal see the backcover.

- from the mind of one man
.....QUite amazing!

THIS ISSUE... er ЕН

I hope Santa was good to all his little
VSUGers and that everybody had enjoyable
Holidays.

This issue we get down to some serious
TS- ing: a couple of members, out of towners,
have sent in some interesting programs for
us to try out; ve finish the 'Sinclair Story'
started last issue; reprinted within is the
article from ‘Electronics’ magazine
mentioned by Harvey about Anamartic,
Clive's company involved in the Wafer Scale
project; Harvey returns with an article
analyzing the price of DRAM over the last
while. If space permits maybe a reprint or

two from the exchange file. Enjoy!
ЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖ

ONE MOMENT PLEASE

Queries have been received concerning the
program listings from last issue. A member
reports that neither the ‘Line Trace’ nor the
‘Dominoes’ program would RUN properly.
These, and other programs, were
photocopied (stolen) directly from back
issues of Sinclair User magazine. Time
constaints do not allow the pretesting of
programs listed in this newsletter as it is
assumed that listings taken in whole from
magazines will work properly. Please let the
editor know if any problems of this type
become apparent in the future. To make
sure this problem does not occur with
member submitted listings, please submit
programs on tape so that they may be tested
at this end before being printed in the
newsletter. As we have all experienced, it is
а real *$%`?@*&$ to type in a long listing
then not have it work properly.

Another member asked if Bill Harmer's
program was missing line 6. Don't fret - this
ine was left out intentionally.

“hank you.

ЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖЖ

2

BITS&PIECES............ nt

..the South Bay TS Computer Club n/i
reports a member uses his TS1000 as part of
a setup to record earthquakes and print the
results on a 2040 printer and yes he did
record the recent San Fran quake. Another
of their members uses his 1000, running as
a portable from a car battery, to run very
complex programs for his model rocketry
hobby at the launch site.

...a couple of user group n/ls report the
expected appearance of a new computer
from Clive's Cambridge Computing.
Apparently Clive has finally caved in to the
inevitable - the new machine will be a 3.5 Ib
portable, MS-Dos, 3.5 inch disk and your
choice of a 20 or 40 meg hard drive. Could
be interesting as the lightest portable with a
DD comes in around 5 Ibs. Look for it around
Apr-May.

...also reported elsewhere is that the parent
company of the North American distributor
for Cambridge Computing's Z88 has fallen on
hard times and has retreated behind the
American Bankrutcy laws. This has resulted
in the Z88 importer, Cambridge North
America, doing the same thing. RMG says he
will not carry, or support the Z88 after the
last one he has in stock is sold. Only Sharp's
is still importing and supporting Z88s.

ыы

се /8 inute
-by your ‘umble scribe

This was another one of them thar
meetings! At 19:30 there were 14 people
present. At 19:40 Gerd finished eating (and
Mario finished flacking him) and Gerd
announced that Glenn Read has been very
busy lately and again wasn't able to make
the meeting. Glenn also has the spectre to
face of the Squamish highway at night. Gerd
then described the pleasures of fighting Quill
trying to get columnar data printed.

There was a discussion of the various
methods by which we might make lists of

the library software available. Harvey
suggested a library disk and Harry Slot
mentioned that a paper listing had already
been promised to out-of-town folk.

The VP & Publisher, Rusty Townsend,
was absent. Ditto the Editor & Treasurer, Rod
Humphreys. (Away in the ‘Windy City’ on
course...Ed.) Harry Slot advises that he is still
the 1000 librarian, sort of. Gerd has the book
library if you want something. Next month
Bill Rutter will bring the whole 2068 library
ind hand it over to Harry.

Mario Vieira disputed the minutes of the
last meeting. Harvey pointed out that there
yere two month's minutes in the last
LXAppeal. Mario was somewhat mollified,
out doubts remain.

Harry Slot stood for the HW SIG and
mentioned that Rusty had been to several
iuctions (swap meets, flea markets) and
jicked up some used TV's for the HW SIG to
nake into monitors. Unfortunately they
yere all of the hot chassis variety...

Harvey stood to present his article on the
price of DRAM 1983-1989. He also
mentioned an article in Electronics about
Anamartic in the UK who have produced a
WaferScale mass storage device; US$11K for
20Megs.

Mario laughed about how Hard Disk
back-up on datasette made the mass-storage
circle complete. Harry Slot warned
everybody present about the dangers of
assuming a single ground was used on
monitor interfaces. Ken Abramson told us of
modifiying the program Run the Country to
the Canadian gestalt. Some students of his
figured out that the way to win was to lay
off all the Civil servants and drive the price
of bread up to about $5 million a loaf (a little
too close to realpolitik).

There followed a burst of strangeness.
Nobody had anything to say. Gerd declared
the meeting adjourned. It was 20:05. We all
laughed & wondered what we were doing
there!

Playing with Electricity

-by Harvey Taylor

-Oct 1/89

The Doc

Well for some unearthly reason I was seized by a compulsion to
know in detail just what has happened to DRAM prices in the last
few years. l am, by the way an inveterate magazine reader & I find
it difficult to throw any of this stuff away. I saw before where
someone on FidoNet had gone throw old mags and generated a data
base of DRAM prices. I decided to do the same.

Below you will find first the raw data which I took from the mags,
and then some graphs I drew from the data. There are several forces

at play generating these prices.

For the last twenty years or so

memory capacity has been quadrupling every 3 years. When new chips
are introduced there is a pattern of sampling, then production
numbers & high prices which come down with economies of scale. As
more chips of a given type are made the yield (ie. the number of

good chips per batch) increases,

which also makes it cheaper for the

manufacturer. Depending upon demand and second sourcing of chips the
price may or may not come down with the manufacturer's costs. Then
there is politics. Some vested interest or other jumps up & down
Screaming that some competitor is cheating & bingo! Tariffs & import
regulations put the price through the roof. The result of the Reagan
administration embargo on memory chips in the spring of 1987 shows 3

up clearly on the graphs.

1
that
to

вак

Jan. 1.25

Fab

Mar. 1.25 4.79
Apr.

May. 1,45 4.09
June 1.95 5.25
July 1.49 5.25
Aug. 1.49 5.27
Supt. 1.56 5.48
Oct. 1.56 5.65
Nov. 1.56 5.97
Dec. 1.56 $.97

1984 16K

64K

did

Wane -150nn -120ns

not
these

any data
have

put
chips

OI)

$200.00.
All of this data is taken from the ads of MicroProcessors
Unlimited in Byte & Computer Shopper. I

the date in the ad. All dollars

266K
150ns -120n

lMed
100ns

100п9- 80по- 60ns

256K
150n5-120ns-100nc- BWns- 60п6

omCcoooco a
=
з
тед

a

200ne 200ns-150ns-120no 100пе- 80ns- ene 100п2

Jan.

Feb. 1.56 5.87 5.99 7.50 69.00

Маг. 1.15 5.87 5.99 7.50 59.98

Apr. 1.21 5.87 5.87 7.50 49.90

May. 1.21 5.44 $.87 6.80 48.99

June

July 1.21 4.99 6.27 8.59 39.97

Aug. 1,2) 4.62 4,87 5.59 35.77

Sept

Oct. 1.21 4.27 4.44 27.97

Nov. 3.56 3.67 3.87 22.47

Deo. 3.07 2.79 3.77 14.99

LE z2222z2222:222222225 i22223z22212222z222222z2c1222000222222222

1965 16K 64K 256K 1Meg

200no 200nc-150ns-120ns 150no-120ns-100nc- ens- 60ns 120га

ттт атттлеттлг zz2222:22z222z22222222-222220222222027
1967 18K вак 56K iMod
200no 200no-150n^- 12006 156ne-120n2- 100net- 80ns- bne 100nc

Jan 1.30 3.04 34.00
Feb. 1,30 3.59 33.00
Маг. 1.30 3.25 30.00
Apr 1.30 4.35 28.50
May 5.35 28.50
June 5.35 27.50
July 4.95 25.00
Aug. 4.95 26.50
Sapt 5.95 32.00
Oct. 5.95 29.50
Nov 5.75 29.50
Dec 5.65 28.00

4Meg chips
just recently dropped from approx.

I believe
US$600.00

in the charts.

Note that the date

US$.

are

64K 256K 1Mog

2¢0ns 200пє - ]5ns- 120ns 150nz-120ns-100ns- 80ns- bna 100ns
4.35 28.50

6.60 36.00

8,40 36.59

9.25 36.50

12.50 36.50

12.95 38.50

11.50 37.50

11.50 38.50

11.78 38.50

11.90 38.50

12.25 3.50

12.25 33.08

zf2z22:22z22222222
256K
1t9ns- -120ns-100ne- 80ns- 60ns

z222222222222222222-2-2222£22222 5227
1Med
100no

64K
200n«-150n5-120ne

Z222z2222222227£7

ттттттзттет

fstrerssfTRRfIC zressfftaztx

The Graphs

+ (1Med 100nn)
* (256K 150nc)
+

.. 4

*

JASONDAFMAMJJ ASONDIFMAM]JASONDJ FMAMJJ ASONDJF MAM
11983 1984 1985 1986 1987

Гезазтзт12222227217222212222222272212117

2222222222222.

ж (* 256K 150ns)
(+ 256K 8@ns)
(o 286K 6@ns)

*
o
4
. о о 00 оо
+ 4 °
ott
weet
LI “зї о
LU
* + o
* oo
++ о
Ы +444
. anne
* 4 stets
serte vetreeteree bette
t: жаз

JASONHDJFMAMJJASONDJFMAMJ JASONDJFMAMJJASONDJFMAMJJ ASOND.JFMAMJJASONDJFMAMJJA
1989

1906 1987 1908

Electronics / December 1989

afer-scale integra-

tion, a technology

that has eluded the

electronics industry

for the past two dec-

ades, will finally be-

come a commercial reality this month,

when Anamartic Ltd. of Cambridge, En-

gland, introduces the first wafer-level
product: a solid-state disk drive.

By taking wafer-scale integration be-

yond the component stage and building

an actual product that has an identifiable
market niche, the three-

growing number of users. OLTP systems
typically handle hundreds of data-base
transactions per second.

Anamartic’s per-megabyte cost stacks
up well compared with other solid-state
offerings. The 40-Mbyte version of Wa-
fer Stack sells for $11,680. Up to four 40-
Mbyte components can be stacked to pro-
vide a 160-Mbyte configuration that costs
$28,160, including controller and SCSI in-
terface. By comparison, DEC's ESE20, a
120-Mbyte solid-state drive used on VAX
and VAXcluster systems, costs about

Wafer Stack's storage, says Jim Porter,
Los Altos, Calif.-based disk-drive indus-
try consultant and publisher of Disk/
Trend Report. Porter adds, however,
that "Anamartic seems to be realistic. It
knows that it has to shoot for targets of
opportunity that need the access advan-
tage that it offers."
In Wafer Stack, two 6-in. wafers form
a module containing 40 Mbytes' worth of
1-Mbit CMOS DRAMs. The wafers are
fabricated by Fujitsu Ltd., which is also
an investor in Anamartic. A controller
board manages data storage

year-old startup hopes to
succeed where other would-
be developers of wafer-scale
integration technology have
failed (see opposite). With
this approach, the company
won't have to wait for third
parties to make use of its
technology, says John Scan-
dalios, vice president of mar-
keting for Anamartic Inc.,
the San Jose, Calif., market-
ing arm of the UK firm.

Although a number of
vendors offer solid-state
disk drives, the market is
ripe for new technology,
says Louise M. Biggs, senior
industry analyst at Data-
quest Inc. of San Jose. Cur-
rent solid-state drives are
built using discrete dynamic
random-access memories on
densely packed printed-cir-
cuit boards.

Wafer-scale technology al-
lows for even denser memo-
ry packing and faster access
times. Among those offering
DRAM-based solid-state
drives are Digital Equip-
ment, EMC, Imperial Tech-
nology, and NEC. Data-
quest projects that the solid-
state drive market, which ac-
counted for just over $100
million in 1986, will near the
$500 million mark in 1992.

Anamartic's Wafer Stack
drive delivers access times
unattainable by other solid-

BY

WAFER-SCALE
INTEGRATION
FINALLY GOES
COMMERCIAL

Stack

BERNARD C. COLE

these 6-in.
wafers in its
solid-state
disk drive,

the Wafer

and transfer, error correc-
tion, and wafer "scrub-
bing"—the process in which
failed memory cells are
purged and replaced, even if
failures occur after the sys-
tem is put in use.

When connected to a host
computer, Wafer Stack emu-
lates a conventional disk
drive, taking up as much
space as an &in. Winchester
unit. Unlike other solid-state
disks, in which individual
memory chips are wired to a
pe board, Anamartic’s wa-
fers are mounted to a carrier
intact. Using the full wafer
eliminates several process-

A British ing steps, including as much
as 90% of the costly wiring
startup, and soldering associated
Я with ре boards.
Anamartic, Anamartic puts 202 dice,
each of which contains a 1-
uses two of

Mbit DRAM, on a wafer,
more than enough to yield 20
Mbytes. The extra dice pro-
vide spares to replace any
failed elements. The 1-Mbit
DRAMs are produced using
а 1.3-рт n-well CMOS pro-
cess. The chips are 13.65 by
4.4 mm’, about 20% larger
than standard devices to ac-
commodate the spare cells.
The DRAMs are organized
as 256-Kword-by-4-bit fast-
page units that require a re-
fresh rate of 52 us.

Each die carries additional

state disks or conventional

magnetic rotating disks: an average of
200 us using a proprietary native-mode
interface and a little less than 1 ms using
a Small Computer Systems Interface, the
company claims. The ESE20, Digital
Equipment Corp.'s solid-state drive, of-
fers a typical access time of 3 ms using an
SCSI interface. Typical access time for
conventional hard-disk SCSI drives is
about 20 ms.

Such speed is certain to catch the atten-
tion of computer designers, especially
those building on-line transaction-pro-
cessing systems now demanded by a

$121,000. A similar system from Imperial
Technology Inc. of El Segundo, Calif.,
costs about $87,000. Imperial has been
selling solid-state drives for about a dec-
ade to a variety of computer makers, in-
cluding Hewlett-Packard Co.

Compared with conventional magnetic
hard-disk storage, solid-state drives and
fast access come at a premium. Storage
on hard disks costs about $5 to $10 per
megabyte; on Wafer Stack, it costs $100
to $200 per megabyte in volume pur-
chases. Customers "really have to want
very fast access” to justify the price of

programmable configura-
tion logie, which Anamartic calls Conlog.
The added logic connects each element to
its four neighboring dice over signal lines
that form logic networks on the wafer.
Using proprietary software, an exter-
nal controller tests each die and pro-
grams the Conlog elements to intercon-
nect the good dice in a single continuous
bidirectional data path. The path takes
the form of a spiral running from the wa-
fer's edge to the center.
Portions of bad dice, but not the entire
DRAM, are routed out of the spiral by the
software. Anamartic partitions each

DRAM into 32 tiles of 32,000 bits each so
that only a failed tile is ever discarded. In
conventional practice, if just one bit of a
1-Mbit DRAM is flawed, the entire device
must be scrapped.

The wafer carrier is a little larger than
the 6-in. wafer, says Anamartic's Scanda-
lios. A silicon nitride layer protects the
top layer of the wafers, over which a sili-
con gel is applied to protect the wire
bonds from moisture. Two wafers are
placed face-to-face in a hermetically
sealed clamshell-like arrangement to
form a 40-Mbyte module.

Each wafer-carrier board includes a
CMOS flash programmable read-only
memory, which holds the map of the Con-
log spiral, including the locations of failed
tiles and spare cells. The flash PROM is
called into play by the controller to access
only good cells after the Conlog has been
established, and to activate spares if ran-
dom errors crop up during scrubbing.

Wafer Stack has already drawn interest
in the computer community. Tandem Com-
puters Inc. of Cupertino, Calif.—like Fu-
jitsu, an investor in Anamartic— "is very
excited about Wafer Stack" based on early
performance results, says Larry Laurich,
vice president of engineering at Tandem.
The company is considering replacing the
hard-disk drives in its OLTP systems with
Wafer Stack, he says.

One of the big attractions for Laurich
is the “no latency/fast access” feature of
Wafer Stack. This feature partitions a
stored data base to boost system
throughput. Laurich says that even
though Wafer Stack storage is signifi-
cantly more expensive than magnetic
mass storage, it still holds the promise of
being at least 50% cheaper than conven-
tionally built solid-state disks.

aes Ltd. may not be the first
company to offer wafer-scale prod-
ucts, but it has a chance to be the first
one to do so successfully.

Anamartic’s game plan is much more
ambitious than those of earlier wafer-
scale developers, including survivors In-
ova Microelectronics Corp. of Campbell,
Calif., and Mosaic Systems Inc. of Fre-
mont, Calif. In fact, both Inova and Mosa-
ic have scaled back their efforts, using
their hard-earned expertise to develop
wafer-scale integration as a silicon-sub-
strate-based packaging technology.

The remaining wafer-scale efforts are
at the universities or are in-house devel-
opment efforts within large vertically in-
tegrated electronics companies and sys-
tems manufacturers. Among them are
AT&T, General Electric, Hitachi, Hughes,
IBM, McDonnell-Douglas, Mitsubishi,
Plessey, and Rockwell.

Mosaic’s first wafer-scale product, in-

HOW ANAMARTIC PULLS IT OFF

TRANSMIT
RECEIVE

—— DATA PATH

[><] DEFECTIVE CONLOG (CONFIGURABLE PROGRAMMABLE LOGIC)

In the Anamartic scheme, each die carries programmable logic (Conlog), which,
under software control, interconnects the good dice in a single bidirectional path.

Anamartic expects to stay ahead of sol-
id-state competitors in terms of price,
says David Hall, chief executive officer
of Anamartic. Because Wafer Stack can
use most of the memory tiles ina DRAM
that has some bad bits, “we can use about
80% of the dice on a wafer,” he says.
“This means that Wafer Stack can be
priced in volume at 60% to 75% of the
price of solid-state disk drives based on
individual chips.” By comparison, conven-
tional semiconductor-manufacturing
techniques may result in a 75% rejection
rate for DRAMs, Hall says.

Although it plans to spend much of the
next year building its position in the 8-in.
solid-state drive market, Anamartic is
considering a number of future direc-

AN ELUSIVE TECHNOLOGY EXERTS ITS PULL

troduced in 1985, was based on a selective
interconnection approzch using silicon as
a hybrid substrate. It used a proprietary
“antifuse” technology based on amor-
phous silicon to form programmable in-
terconnections betweea metalization lay-
ers. It attached ICs to the silicon sub-
strate and wire-bonded them using stan-
dard techniques.

Mosaic targeted its offering at military
and aerospace applications—without
much initial luck. The company lost about
$4 million last year, with sales barely ap-
proaching the $500,006 mark. Now under
new management and with a second
round of financing, Mosaic has altered its
sights, targeting:a number of new high-
growth market segments, including por-

table computers, local-area networking,

and modems. It hopes to push sales to
about $15 million to $20 million by 1991.
Inova also went after military and

aerospace customers in its effort to bring

tions. A first step, says Scandalios, will be
to reduce the size of the subsystem by re-
placing many of the discrete components
used in the controller with a few gate ar-
rays. Slightly redesigning the basic sys-
tem and squaring off the 6-in. wafer
could make a 5.25-in. drive possible. By
using wafer segments, Scandalios says,
3.5-in. solid-state drives are within reach.
And a shift from the present 1.3-4m pro-
cess to a 1.0-um process could boost both
density and speed considerably.

In the long run, Scandalios adds, other
types of memory—indeed any regular ar-
ray of identical components, such as
EPROMs, electrically erasable PROMs,
and static RAMs—are possibilities for
wafer-scale integration. Г

wafer-scale CMOS static random-access
memories into the 32- to 64-Mbit range.
The company based its approach on a pro-
prietary interconnection scheme it calls
Inroute. In this approach, wafers contain-
ing slightly modified standard SRAMs
are fabricated by a foundry; to them In-
ova adds a third proprietary interconnec-
tion level that links good and partially
good dice.

Like Mosaic, Inova has reformulated
its marketing strategy; it is now hoping
to serve the high-speed CMOS SRAM
market by packaging partial wafer slices
in standard memory packages. This mar-
ket is expected to grow from $480 million
in 1987 to almost $2 billion by 1992, ac-
cording to Douglas Mitchell, vice presi-
dent of marketing at Inova. Within 18 to
24 months, he says, the company will be
in production with an 0.8-um process for
building a new generation of 4-Mbit
SRAMs. -В. C. C.

zt Ca fu Ee

ИП
-l
Ti

6 ERR

П s ETT]

ке]
ж
©

[eR CO FU Es н CO (Л {+

тарро p pe
4n
+

uu
г
ЕГП
4
га

шын шош dB mm MEE ПБ
Hn

mm om om
4n
4

~Ji
г
n
~i

cun or
Г"
m
. a4
x= Moo ono)

sma С
AE ++
Att ++

кип

m
nm

ATO p 5" TO E^ fà e be 2 E» pe CL RS

5 CI Ra FU s ©
ino ananas

[ж]
“se
E

200 NEXT М

205 LET O¢= DEI H.C
TS

sXL. X. IX. U. IVI."
= FOR N=1 ТО
GOSUB 1499
LET С (М) =U$
NEXT MN

і
1
2

CLS

PRINT
4UHBERZS
PRINT
PRIHT
PRINT
PRINT

_PRINT

SSES

DRE

X E Fa MEL UD

=
-

2 re

38 PRINT ,,TRB

340 PRINT » THE
ROMAN NUK BERS 2

(ә)
z B En ma

Leg PRINT HT 19,8; "ENTE

cnt
COP CIL CIL CL CIL CIL CIL TO CIT Fes CIL CIL CIE Ps +}
74 £n Cr] £e Ca TO E» C) E

Ш

^

сл
1

+

74 CI TI Ea c Ca Ca Ci -d C]

£n oen cn cn en on qn en e cn Tm OT Cn On on Л

LET S£z"HDCLXUI

ES ES | M 24 74 Cni cn cn cn cn cm en

E

8
Sg
Б
zi
ii
e
G
1
1
1
2
2
4
N
5
[=]
Е
7
7
a
E
a
1
3
4
5
5

£5 3 (C a a 020 (C93 a (Л C CI] Ca С) ССС 089 C9 СЛ Ga tC UT] a

"m

M

“nor

UAL NS
(8 OR Ni»INT N THEN GOT

! LET R$z""

LET КЕТ

LET HTzH-FíFLI)

IF NT: THEN GoTo
FOR L=LEN CS EFL

Е
ТРЕ КЛУБУ" "U^ THEN GOT

NEST L
LET RS=SRS+CSFL, TO L?
LET NzNT

GOTO sea

LET FL=FL+1

IF FL:=13 THEN GoTo S
PRINT AT 18, (31-LEN R

PRINT HT
SLOW
PAUSE 4Е4
GOTO 599

PRINT AT 10,3; "VALUE BEYOND
GRAM LIMITS"

PRINT ,,THB 1; "ENTER A
WHOLE NUMBER"

PRINT AT 28,3;K8

PAUSE 454

GOTO 399

REM

REH ‚Б онан то Швавтс

LET Nsan"

INPUT N

FAST

CLS

IF H£z"'" THEN RETURN
LET TL= -ü

^q d

єк” Gn
Cho аа

1
@

LET UzTíPC,CC) i aoe
IF Ws@ THEN GoTo 18858 DNA
LET TLsTL4 1120
IF CC»zPC THEN 1136
IF Ltaci< e 1140
LET F-1 1156
LET Бшш СР) 1499
cara 920 1566
LET Е=@ a

ET PL=c 1510
EEF oc =P i326 GOT fae
LET РС= 1530 LET UECDE (DE TO Р2-1}
LET D=D is48 LET РӘ=Р2+1 H
IF РГЕН "NS THEH Gi 275 isse LET Pi-P2
GoTo 750 1566 RETURN

599 REM .READ ON ROHAN NUMERALS

LET RS=SSTRE TL Bs 16 RA
PRINT AT 16, (33 H RRSP SER 1700 CLS

1710 PRINT AT 3,9: “ROMAN NUMBERS
PRINT AT 20.3:K8 - 22, "ROHAN HLUHBERZ

PAUSE 464 1720 PRINT ,,"THE MODERN
GOTO 61g EC ALLOWS | A TUO-LETTER C
iege PRINT AT 18,8; "INVALID CHAR OMS "a WHICH MAG JD
ACTER FOUND EET PEL Бад i988 PRINT. бае Же
1010 PRINT j| Б; SE ONLY H 7380 PRINT ,,"IM THESE CHSES,THE
,D,C,L,X, uU, I^ SMALLER", “MEMBER IN THE REUÉBSE
1815 PRINT AT 20,3;K& PAIR IS", "SUBTRACTED INSTEAD OF
1825 bora етө" 1540 PRINT
B25 GOTO 6195 i740 PRIN 12 RHLS0,H LETTER CAH
1030 PRINT AT 10,0;"TOO MANY ";F BE USED UP","TO THREE TIHEZ IN s
$: 85 IN A ROUT, (MAXIMUM з)“ UCCESSION.","""H"" IS THE EXCEPT
1835 PRINT AT га, Ake ION.IT CAN БЕ", "USED ANY NUMBER
4848 PAUSE 4E4 OF TIHES IN", "ÜNBROKEN SUCCERSSIO
1@45 СОТО 619 ч."
1858 PRINT AT 10,8;"IHMURLID CHAR i778 PRINT HT г@,з;к®
ACTER SEQUENCE -" 1780 PAUSE 4Е4
1660 PRINT o N$ 1790 GOTO 27€
1070 IF O=1 THEN GOTO 1874 1999 REM .CLOSURE
1671 FOR A=1 TO D-1 2606 CLS
1872 PRINT " "; 2618 PRINT AT 18,14; "BYE": AT 21
1873 NEXT A @; “END PROGRAN" i | 25
1074 PRINT "+" 2826 GOTO 2802
1875 PRINT AT Z86,3;K£& 2100 REM To TEST PROGRRH
1080 PAUSE 4E4 21805 REM ENTER 1989 INTO
1255 GOTO 813 2106 REM ARABIC/ROHA CONVERSION
190 LET ОЁ=@ 2110 REM ENTER HMCHLXXXIX INTO -
AE QO«LEM R£-1:LEH Gs THEN Sir БЕМ Plea areata cba CONUERS TON
RM 21: Ч о HUMBER R IRNS T
2116 REM OTHER SER RETURNS ЕНЕ

m
SOLDERING TIF...

Does your soldering iron overheat while
on standby ?...I am sure it does, unless
you have one of those very expensive
temperature controled stations.

A very inexpensive solution is the HI-LO
FEED-THRU DIMMER, made by LEVITON, rated
at 300 watts and costing only $4.99.

It is nothing more than one of those
little switches that go right оп the
lamp cord, with a built in diode.

à rotary knob gives in sequence, GFF,
HALF and FULL power. As there are two
full cicles for a full turn of the knob,
I found it convenient to mark the full
power position with a small round file.

8 Instalation of the device takes a just a

couple of minutes.
Marcio Vieira

CURVE SKETCHING

by Alvin Albrecht
Type in the listing (omitting
rem statements and/or the
instructions -line 5 and lines
1000 and up- if you want to
shorten the typing job) and, if
you have СЕ Type, do a БЕ type
check with the Ck check output
following the program listing.
If you deleted the rem
statements and/or the
instructions the lines that
contained a rem statement will

have different CE outputs.

Before you save this program you
should know that there is a bug
in it. I didn*t realize this
until after it was printed and
the bug was so easy to correct
that instead of reprinting the

entire listing, I am providing
the changes:

iio GO TO 185

ба PLOT x/xscale-ti27,
vy/yscalet87
IF РЕЕК ZZ729-11 THEN
IF PEEK 22736-256XFEEK
THEN GO TO 185

озгулуу= рощ
act F Vm 1 a

257 IF PEEK 237239-11 THEN
Ga TO 184
250 BEEF .O005,20:IF PEEK

237Zz9-10 THEN GO TO 185

Now save with: SAVE "F(OX)SKETCH"
LINE 1

Run the program and read the
instructions provided (If you
omitted the instructions when
you typed in the program, read
them from the program listing &
line 1900). The following are

some examples to make you better
acquainted with how to use Curve

Sketcher.

Set Range-t/- 2, Domain--t/-
2XFI, Function: f(x)=SIN X
(enter SIN as a keyword. ie-
you normally would from basic)

as

9

three
and 10
screen.

Next graph the function
times with zteps-4dO,20,
without clearing the

The graphs zhould look
inaccurate (mildly put? but
there should be an improvement
as the step is lowered.

Now. clear the screen, set
step-Z, and graph the function.
This is what the graph of SIN

should look like. A step of 3 is
ideal for this function because
it is not too slow and it is

accurate.

It makes sense that if the step
is lowered further an even more
accurate graph would be produced
(at the expense of a longer time
to graph it) and this is true,
but we run into different
problems. Clear the screen
again and set step=9.5 and graph
the function. A more accurate
graph is produced but it looks
fuzzy.

know the
# GO =SIN
values of x SIN
clear the

want to
equation

Suppose we

roots of the
x (ie-for what
HEO) a Set step-i,
screen and graph the function
again. Choose the "Coords"
option from the menu by pressing
"2". SIN «=O when the graph
crosses the x-axis. se the
arrow keys (5 to 8 will move one
pixel, CAFS + 5 to 8 will move
8 pixels at a time) position
the flashing dot on

to

one of the points where the
graph crosses the axis. The
coordinates of that point are

given in the bottom left corner.

One of the roots is the

coordinate. If you have a
calculator handy you can check
this by entering and take the
SIN of that (don?^t forget the
2068 works in radians) and the

answer should come up very close

to zero. The slight inaccuracy
is due to the fact that each
pixel width on the screen
represents several x-values. The

range of values is determined

by the horizontal scale. You
must also be careful in
selecting a suitable domain and
range. Clear the screen again
and try graphing this:

Domain=+/- 5, Range=+/- 3X,

Ж, Function:

TO) =¢x4+1)3 / (sxx)
Remember what this looks like,
clear the screen, and graph the
function with one change:
Range--/- 7. Notice the extra
part of the curve? It wasn’t
Plotted before because the
section of garaphpaper we used
the first time had only the
x-axis from -5 to +5 and the
y-axis from +3 to -3. The extra
Parabolla’s vertex is at —5.2 on
the y-axis. This feature can be
used to get close-ups of
sections of curves for detail.
Did you notice that the hump on
the leftmost part of the graph
was harder to see on the second
graph?
Опе last example: Range=+/- 2,
Domain=+/- 2, Step=5
Function: f¢(x)=LN x. Clear the
screen and graph this. Why is it
beeping at you? It's
encountering plenty of
mathematical errors. The LN
function is undefined for

negative numbers but the program
is attempting to calculate LN

for —24£x42 (the domain?. The
beeping is over fairly quickly
and the graphing didn’t take
that long, but far more
complicated expressions the
computer may be wasting its time
calculating futilely the values

of a function at undefined
points and this will cause a
considerable delay in the

graphing. To avoid this problem,

we change Lstart. Lstart tells
the computer from what FIXEL toa
start graphing the function
from. When Lstart--127 (the
default value) it starts

graphing from the left edge

the screen. If we set Lstart-o
it will graph from the centre
(origin?) of the graph. So to
only graph the function for
positive values of s set
Lstart-o. Do this, clear the
screen, regraph the function,
and verify that there was no
beeping.

What if we want to start the

graphing at a specific VALUE OF

X. eg- we want to see the graph
of LN x only where «30.73. Set
Letart=0.73/xscale. Easy enough.
(Try it if you like

after clearing the screen and
use Coords to verify the first
point plotted has an x

Q.73).

coordinate of close to

That about
narrative.

wraps up my

Hope you enjoy it.

A шч o] v ощ G беги, A n^ AML: R
[лб 4 xm ж О T-L " © н а чм т ANZI m а ст
nu О mm x ms н ul ы... xnl с Фе a E
c 0 ж СЕ qi CD t wil о іл EGLI н Oc px ocu H
E GW AD nim-« о WR bz ы. U e HZ Lh] ч* A WOR © A ou) dg
U + g^ +00 сч +O 5 Ф Ән ~- 00 ni чир nm Ie
о YW O Еш pn: OY KY Q 1 a B HKA etm Ож 0л rx ae
H ШФ = OY ok OO nix © а 00 c Шш > HO U пан 0 - U t
E d! ^ HM Ge AM D — шіЛа. 2ч - (Л - XO A “Van ^ - ^ ч
C @ 9+ muy oo T AZ Qni (ЛЕ н -oN у TUS ү - м @
(9 t сс OMA шї. OF аш ш ü eI U - WX A +++ A - 8 E
OM mP* U Ht Ht D ш on c i ЕМЫ + ^ v Um -GLuolt) -D о
L n. 0..4) ow LS oO 0 єн @ n.o 400 0G 2h G 2704 8 Q
[t OF on О ж OA a х T D Q P9 E ^ ш "m Zv ty - [n
4 Q Ol O^ 0-7 фо HOW Q^ A (Л sy n. 0 Ө oZ н OLLHA O н 3
ni Ф ACH 5 ah RO Wy R > Ф ано Li һ. v 07044) Q
ü 20 eU mv elt AD IM O HA ш x n Oh он RY E 2 b н 2
ul шл evo O+ шо бийч ш dv x x кс (D a q a © TAL: UGA T ©
X rj Wt non wr (ino ‘AS оо KANU «ня LOG Ro EAXUI-- NMS h
| 25 * POR a pwr a a dun р O> н зч д. ї c3 3 © Um Paeh Gmüd-e rus go
X C ll ^ Oli QA *üu t| ^ FH vA iibi QU LON e~ Q 3tÜ)40— OHNA U~ Wt Oy p
I о я peu UIAM: я immrz q x ANZ LO A0 O Z a z: M tR vm HM e
X E. 1э+аюніодчллайі Dru з н2 ofr ай -Neke (E D> Bit DOme? ноя LDO 0900 ш
лс ^ ла u uui гє Б дё) nu O sD zh: лн2 1050 utz-ut-rn
HELEH тл + озн ъ= нны Гл Qu) ненщ см 1+ 052 eun LEL
[ИПИ Ory Q^- WO O^ йш шн. шОшшщы bj OR Uhm- X 0004 @ +ULOOr СОС
04, 1 o .-i 05 * AIO OF хири wk HUH, Sex: gv (9000, Oy m4 ouarn-
H O0 amx o Ож O Weil il n: gs с x

Snes SEND м: DIENS м UNeANS NOSOS N-O EZNmMMEINT O+ S2u tv ^Bnzeas-unzatrdo)
MOSSO ALAC «1 (ПОО ql qudunuuto qOrOG бд. ANMNSSLAD alm: A МЕЧ IU CUL mo) 70)0)
PUM pa FUR осу e T PAI PrP po prt uu) NEXE . AMNU: л (OT + UI

CURVE SKETCHING

1 REM

Dec 1989 Albrecht

| ^O OHA on © Z WOE ї ^R АСФ
S pH zd an ^ mU нили = Wz OM шщ
~ e ix N @& 2 m HY WU х ош BE. m (л
M iow UG e ш н | (Л Хо ANA T ou IN mc
^ Фр Q Fae cn Dn. (0 X om U ms ахх л 3c „шшс m q
8 cU X Iv о. ш 2 н au ч aX a vg M чос н +0 WU A
QS оч qc Xt P How os X I vp dd 4) (UH Go
© pü в (щл x N 420 | pH 203 uz oo 20-2 “OZ
ч YUH 3 02 n nz [os] Od x nO NATO 0 a ON Wm) ww
ж Coy 02 fu т гг OQ 4 0 0] cau ч ЫШЫ AUN rigor uer
X ^u. ut О.- ч U 3 м 4 na ak, ч 07 лн чус нмин 0. н
w н Qo 23 X as e^ О (Л зох Keon w aujo mo С)
ch. Z O all .- Ф ч нх TO - 5 к ш 00 Ммм ыч
ын us xr NON с © ~ а. WN — 09+ AONI шщ Иа Mmm = U
Z n X WE C U~ ш AO. 0 sx + азе ом2-ч+0 1 1 nO QO ufu -I Oi
mi > Шш a GO I ae QO О 4H Q чос n4. ч _ ш g) m0) On
01 e || x goo c 40 An в zu c Zo Dn On Oo m ELON mxnm © m
7^ OIA - ы RRO e GOR om LJ U а WUBP OHAD J X ш GA C Male UN
Ох 900 8: 009 T G дла X X IH щлон-ч 000-4 ci (Л @ PA miu«m 1)
uv 84 GG Orr | pX JD O HX © о нач0о0-2 0л а р ш QUJ32 па. in u
Oh At NO з UGe р N*O TU | oO + -10ZORQN Ww р nn | COV * = ын ©
m OL] + mu жо оО mnt D H 4X d HDI 1и HD- с mv C2 О XUwux oxxo
м2 Mnu m ш Б e HHI l I (Л о owu x Онон 0х0 о ч к OOu- WNOL! bie
bh, 2X DO а ано од. хын -3m 29250 О чоо0хочочн 0 х nm 2шоа WUNA 2щ
ш л « ur ш2 0 І z^ аб a3 oh пч IZA. -w шщ Q.* xD. a, 0
ми. He ILL =O BYR XH NECE CU кк - 0- PMOEEER рули x юм WZ н
OW Oh OOWWW2& Xx iu ИОА NOW, Qu. Ni Иш! тщ Ша хь. шши.
BO DOIN OUOREMOMs* 7, OA јн MeOH wmm KEE) Cn RH = MI.O
N GO - о ES (Ü hl „++ ~ Mm HE 0
UM МӘНӘ ломе «сл mc 0605-7 ene ШЧО uU

SWAHUNMNGSSy 08 ӘӘ. eAnufunu-J-amomunuu NN Quz -»xP"«eogoegyru
9 = | tet {4 Od eee Odie О Matti KAZ (0 | ate enm
8 y л ш z 19 cm Ss dim corny) Ty}

nun NOAS ::
MUM OU

OTO? LOZCLOLOCOLO
Er » (x

Continued next issue.

11

Continued from last issue.

= The Sinclair Story

2s predicted:

ways been based on being first

with products, often aimed at a
market that didn't know it existed. By
1979 there was a well established
‘personal computer’ market. Com-
modore had launched its £700 PET
home computer the previous year.
Apple and Tandy were also well-
known in the field. These machines
were found variously in laboratories,
and commercial and teaching estab-
lishments; not many people had a
computer at home.

Sinclair decided that he would have
to offer a product with all the essential
features but at a greatly reduced price.
In May 1979 The Financial Times
"Personal com-

puters will become steadily
cheaper and their price could

Gree SUCCESS had al-

E!

E

12

drop to around £100 within five
years." Typically, Sinclair decided to
do it in a few months.!

The ZX80 — the world's smallest
and cheapest computer — was laun-
ched at an exhibition in Wembley at
the end of January 1980. It measured
9" x 7" and cost £99.95, or £79 in kit
form.

In order to keep the price low the
designers had to introduce some radic-
al ideas to reduce vastly the number of
components. The biggest saving was
the use of a domestic television set as a
Screen and a cassette player as a
program and data store. The machine
had a Z80A microprocessor which was
supplied by Nippon Electric; a large
ROM, which contained a 4K-byte
specially written Basic interpreter, the
character set and monitor; and the
interfacing circuitry.

The ZX80 was very much aimed at
the person in the street wanting to
know something about programming
computers. Sinclair was convinced
that people could be persuaded to buy

the ZX80 but how to persuade them
was the problem. The image of the
computer at that time was somewhat
Big Brother; clinical, air-conditioned
surroundings; huge cabinets with reels
of magnetic tape whirring to and fro.
How would people relate such a
frightening piece of equipment to the
ZX80? Why would they want to buy it
for the home? Why would they want
to buy it at all?

No one need have worried. The
ZX80 was an immediate success; ten
orders were placed at the exhibition in
the first five minutes. The office in
King's Parade was suddenly inun-
dated with cheques; the switchboard
was permanently jammed. Nobody
had expected quite such a response
and there was total chaos. Clive’s
immediate problem was to ensure that
the company could cope efficiently
both with the administration, and
with the production of the ZX80.

Sinclair wanted to sell the ZX80 in
the United States, although he did not
expect to find an enormous market

Б,

A

there because of the strength of the
competition in the. home computer
field. However, a few weeks before
the launch of the ZX80 in the UK he
took it to the Las Vegas Consumer
Electronics Show, and at the same
time met Nigel Searle in Boston.
Within a few days Searle had a new
job, a new apartment and an office in
Boston. He sold the ZX80 and later
the ZX81 in the States from that office
by mail order until early 1982.

Sinclair Research expanded rapidly;
by September 1980, over 20,000
ZX80s had been soid. Clive Sinclair
was determined to keep the company
to a manageable size; he was all too
aware of the need to try to learn from
previous mistakes. Bringing manufac-
turing in-house in the days of Sinclair
Radionics had seemed an excellent
idea at the time, but the number of
people they had had to make redun-
dant had hurt him deeply.

By this time there were 12 em-
ployees at the King's Parade offices in
Cambridge, six engineers still working
at The Mill in St Ives, and Nigel
Searle in Boston. To make sure that

the company didn't grow too fast
Sinclair had subcontracted all manu-
facturing. To begin with, production
was done locally in St Ives by Tek
Electronics. Components were gener-
ally of a much higher standard than
they had been during the Black Watch
fiasco, so there was less reason to
manufacture products in-house. Even-
tually, as more and more were pro-
duced, the computers were made by
Timex in Dundee; it is a testimony to
all concerned that the return rate on
the ZX80 was only one per cent.
Although the machine was so popu-
lar and sold so well, this was largely
because it had no competitors. In fact
it did have some drawbacks such as
the lack of floating point arithmetic, a
capacity of only five digits and an
inability to handle separate files on its
cassettes. The touch-sensitive — or
sometimes touch-insensitive — key-
board was unpoular with users too.
But in spite of those shortcomings,
the ZX80 had opened a new market
sector which exceeded Sinclair's wild-

—

SINCLAIR USER January 1986

est dreams, so who was going to
complain too loudly? In September
1980, the company launched a 16K
RAM pack — an extra plug-in mem-
ory — to attach to the edge-connector
at the back of the machine. There will
be many who remember the well-
known RAM pack problem whereby a
slight breeze could upset the connec-
tion and an evening's work would be
lost. Thank heavens for Blu-Tack.

The ZX81 was launched in March
1981. It contained a new chip, de-
signed by Sinclair Research and
manufactured by Ferranti — the
world leader in uncommitted logic
arrays — standard chips which can be
adapted to a user’s requirements at the
last stage of production. The new chip
replaced 18 chips in the ZX80 and the
machine now retailed at £69.95 or
£49.95 in kit form. Sinclair also
offered an add-on ROM to convert the
ZX80 to the ZX81.

The ZX81 had a floating decimal
point and scientific functions. It came
inasturdy black case and, if you used a

colour TV, would produce black char-
acters on a restful green background.
It was a vast improvement on the
ZX80. Sinclair also announced that he
would be launching a small printer to
work with the ZX81 later in the year.

Now that he had an improved
machine and the promise of a printer,
Sinclair decided to fight back at the
government's scheme by offering his
own half-price deal. Schools could buy

MA continued on page 88

continued from page 87

a package of a ZX81 and a 16K RAM
pack for £60; and he further promised
that they would be able to buy the ZX
Printer at half price when it was
launched. That made the total cost of
system £90, while under the govern-
ment scheme the minimum a school
could pay if it bought an 'approved
system' was £130. About 2300 schools
purchased the Sinclair package.

The ZX81 received a very sym-
pathetic review from David Tebbitt in
Personal Computer World in which he
keeps referring to *Uncle Clive'. On
the other hand: “Sinclair has been a
bit cheeky in his advertisements.
Under a column entitled *New, im-
proved features', he proceeds to men-
tin three things that were included in
the ZX80 when it was launched over a
year ago!"

The ZX Printer was eventually
launched in November 1981 at

£49.95. Designed for the ZX81, it
could also be used with the ZX80 with
an 8K ROM. It was a very compact
little printer using a special metallised
paper, and would print 32 characters .

n and towards the end O
the year was granted exclusive dis-
tribution rights for the ZX81 in Japan.
Mitsui was one of Japan's main impor-
ters of British goods, the range includ-
ing Jaguar cars and Burberry rain-
coats. They planned to market the

Kesea

ZX81 by mail order at about £90 апа,

Sir Clive dons his running shorts

3 = EA fia
Nigel Searle in Boston
aimed at selling 20,000 computers
during the first yest; there were no
competitors.

By the end of January 1982,
300,000 ZX81s had been sold world-
wide. In the USA Sinclair was selling
15,000 personal computers a month
by mail order; American Express was
selling thousands to a potential ten
million customers. ‘Then Timex was
granted a licence to market both
current and future Sinclair personal
computer products in the US from
mid-1982. They paid Sinclair a five
per cent royalty for sales and bought
the right to use the Sinclair name in
the US. .

In Britain, Sinclair signed an agree-
ment to sell the ZX81 through the
branching-out stationers and booksel-

П fous retail Outfets were just опе
of the ways in which the home compu-
ter created jobs. By February 1982
production of ZX81s was running at

about half a million zachines a year ,

and the company had a turnover of
£30M compared to £4.65M in the year
ended March 1981.

One of the interesting side-effects of
the ZX80 and ZX81 was the number
of cottage industries that sprang up
because of them, producing software,
peripherals and publications. A ZX80
Users’ Club had been formed before
the ZX81 was launched; SYNC
Magazine appeared in January 1981 to
cater for ZX81 users; Learning Basic
with your Sinclair ZX80 by Robin
Norman, published by Newnes in
early 1981, was one of the first books
to develop Basic programming techni-
ques on the home computer.

Hundreds of small operations
started to sell programs, books, extra
memory, printers, sound generators
and add-on keyboards for use with the
ZX81. In January 1982 one Mike
Johnston organised a fair for com-

panies selling products for the Sinclair
computers. Nearly 10,000 people
turned up at Central Hall, Westmins-
ter, which has a capacity for only a few
hundred; the police had to be called to
control the crowds; 70 exhibitors took
huge sums of money.

Both the ZX80 and ZX81 had been
produced as learning machines; for
the person wanting to find out about
computer programming. Once people
knew what they were doing they
wanted a more powerful machine, and
at first they had to turn to manufactur-
ers other than Sinclair Research to
find them.

Sinclair’s philosophy — at least in
retrospect — was to prepare the world
for universal computer ownership in
easy stages. Over 50,000 ZX80s had
been sold, and more than six times as
many ZX81s. As the market matured,
the engineers were working away at
the ZX82 (codename) which was laun-
ched as the ZX Spectrum in April
1982. The hardware was designed by
Richard Altwasser, who later formed
his own company, Cantab, and fell by
the wayside in an attempt to market a

sions: the 16K sold for £125 and the
48K for £175. For those who prefer-

red to work up in easy stages, an extra
pack to increase the memory of the
cheaper machine was available for
£60.

In many ways the Spectrum was
altogether a ‘better’? machine than
either the ZX80 or ZX81, although
some said its predecessor the ZX81
was superior when it came to finding
out how computers actually work. Its
chief advantages over the ZX81 were
‘eight-colour graphics capability,
sound generator, high-resolution
graphics — smaller dots on the screen —
and many other features, including

continued on page 90

continued from page 88
the facility to support separate data
files."

At last, Sinclair Research was
notionally able to compete with the
BBC Micro and other personal com-
puters; the figures in the table pub-
lished in the ZX Spectrum leaflet were
impressive. The ZX81 had been com-
peting against the Acorn Atom; it
could never have stood up against the
BBC model A, the current Acorn
competitor when the Spectrum came
out. The Spectrum had a more versa-
tile Sinclair Basic than the previous
two machines; an improved keyboard
replaced the unpopular — though
cheap — touch-sensitive keyboard; it
was able to generate and display

IDE 4i

D цан а ü
rubber pad over a ZX81-type mem-
brane keyboard, and which had a most
peculiar feel to it.

The Spectrum was the cheapest
home computer to produce colour
graphics but the reviewer complained
of the lack of facilities and ‘found that
the borders tend to wriggle in an
irritating way’. It also had a small
built-in loudspeaker which generated
bleeps ‘acceptable for games, but not
much more’. And that, to Sinclair’s
disappointment, was about all the
Spectrum was generall
used for. The

One-piece їп

/

элс pi EES

Sinclair's headquarters in Cambridge

tone of the review was set in the first

paragraph:
“After using it, however, I find
Sinclair’s claim that it is the most
powerful computer under £500 un-
sustainable. Compared to more
powerful machines, it is slow, its
colour graphics are disappointing,

its Basic limited and its keyboard

meant that parents were prepared to
buy them to give their children ‘a good
start in life’.

The place of the computer in the

iA nal е - |
жч

The launch of the Spectrum

meet the delivery date,” he said.
Ronald Harris of Swansea said he had
writing since the end of Мау. “1

home was reinforced by the meagre
provision in schools, where there was
often only one machine between 30
pupils and thus insufficient opportun-
ity for everyone to practise. What
better solution than a computer at
home?

But Sinclair observed another
dimension: "The interesting thing is
that as well as children being expert at
programming, there is another expert
group taking to it like ducks to water
— retired people. The concept of it
being peculiarly suitable to the young
mind is perhaps wrong — it’s the
mind that’s free of everyday burdens.
The retired person with some time to
spare can take to it wonderfully and
it’s giving a lot of people a new interest

ШК

П Ш its aK mem
was capable of playing very sophisti-
cated games and there were companies
starting up solely to produce them —
often run by very young people who
had learnt programming at school or
from magazines.

In February 1983, WH Smith, who
had been the Spectrum’s biggest dis-
tributor, was joined by Boots, Currys,
Greens — Debenham’s in-store sub-
sidiary — and John Menzies as Sinc-
lair pioneered a change in the High
Street. Many other stores such as John
Lewis and the House of Fraser were
supplied by Sinclair’s UK distributor,
Prism Micros. 200,000 Spectrums had
now been sold by mail order, and by
Easter 12-15,000 Spectrums were
being sold per week in the UK. The
Spectrum had also been launched in
more than 30 countries worldwide.

You couldn’t walk into WH Smith
on a Saturday without being faced
with shelves of software and mobiles
and whizz-kids playing on the compu-
ters. What sort of computer you had
became an important factor in play-
ground status.

And where has it all led? Computer

awareness has been generally raised; `

the dust has settled, much of it on the
home computers, leaving a hard core
of enthusiasts. The market is satu-
rated; the craze is over. The computer
is settling into a serious niche compa-
rable with ham radio; the days of the
CB computer are surely over.

15

2z»mooo0o-m-O»7o

alts

The Vancouver Sinclair Users Group has been in existence since
1982. We are a support group for the owners and users of all
SINCLAIR and TIMEX computers.

a А

Prez:- Gerd Breunung PH#(604) 931-5509
V/Prez & N/L Publisher:- Rusty Townsend
Scribe:- Harvey Taylor

Treaz. & N/L Editor:- Rod Humphreys

Our membership dues are only $15.00/year and may be sent to the
Treasurer:

Rod Humphreys
2006 Highview Place
Port Moody, B.C., V3H 1N5

Members of VSUG receive a monthly issue of ZXAppeal - our
newsletter.

ZXAppeal accepts advertising. Our **PREPAID** rates аге:

$10.00 — full page
$8.00 — 1/2 page
$5.00 — 1/4 page

ZXAppeal is distributed to approx 30 other T/Sinclair User Groups
throughout North America as well as overseas via the NETWORK.
NETWORK correspondence may be directed to the Editor at the
above address.

Copyright of all articles appearing in ZXAppeal is retained by the
author with the understanding that other T/S User Groups may
reprint any article appearing in ZXAppeal provided credit is given to
the author and VSUG.